Conformal Object Detection by Sequential Risk Control

2025-05-29

AI Summary

To address the lack of statistical reliability guarantees in object detection models, this paper formally introduces the task of Conformal Object Detection (COD), the first to extend conformal prediction beyond single-output settings. We propose Sequenced Conformal Risk Control (SeqCRC), a novel framework enabling provably valid coverage guarantees for joint bounding-box and class predictions. Methodologically, we generalize conformal risk control to sequential, two-parameter decision spaces; design detection-specific loss functions and prediction set construction protocols; and establish the first COD benchmark with an open-source toolkit. Experiments on PASCAL VOC and COCO demonstrate that SeqCRC achieves a practical trade-off among guaranteed coverage, localization accuracy, and computational efficiency—under strict statistical validity.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Constraint Satisfaction and Optimization: Constraint Optimization
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Recent advances in object detectors have led to their adoption for industrial uses. However, their deployment in critical applications is hindered by the inherent lack of reliability of neural networks and the complex structure of object detection models. To address these challenges, we turn to Conformal Prediction, a post-hoc procedure which offers statistical guarantees that are valid for any dataset size, without requiring prior knowledge on the model or data distribution. Our contribution is manifold: first, we formally define the problem of Conformal Object Detection (COD) and introduce a novel method, Sequential Conformal Risk Control (SeqCRC), that extends the statistical guarantees of Conformal Risk Control (CRC) to two sequential tasks with two parameters, as required in the COD setting. Then, we propose loss functions and prediction sets suited to applying CRC to different applications and certification requirements. Finally, we present a conformal toolkit, enabling replication and further exploration of our methods. Using this toolkit, we perform extensive experiments, yielding a benchmark that validates the investigated methods and emphasizes trade-offs and other practical consequences.

Problem

• Ensuring reliable object detection in critical applications
• Extending conformal prediction to sequential risk control
• Providing adaptable loss functions for diverse certification needs

Innovation

• Sequential Conformal Risk Control method
• Loss functions for diverse applications
• Conformal toolkit for replication